CV

Education

• Ph.D. in Computational Engineering, Durham University, 2023
  • Thesis: Automatic Extraction of Ordinary Differential Equations from Data: Sparse Regression Tools for System Identification
• M.S. in Data Science, Northwestern University, 2019
• B.A. in Statistics & Mathematics, Wittenberg University, 2016

Work experience

• April 2024 - Present: Postdoctoral Research Scientist
  • University of Hawaiʻi at Mānoa, Honolulu, HI
  • Developed Physics-Informed Neural Network framework for 4D-Variational Data Assimilation achieving 6.5-fold computational speedup over traditional numerical methods
  • Submitted manuscript to Journal of Advances in Modeling Earth Systems
  • Leading a Simons Foundation Targeted Award–funded workshop on machine learning methods for time-series data (2026)
  • Contributed to development of external research funding proposals focused on scientific modeling and data-driven methods
  • Supervisor: Dr. Brian Powell
• September 2017 - September 2019: Associate Business Analyst
  • Medline Industries, Northfield, IL
  • Reduced rebate inconsistencies by 30% through systematic analysis
  • Established relationships with stakeholders to resolve contract discrepancies
• January 2016 - May 2016: Data and Assessment Assistant
  • Springfield Promise Neighborhood, Springfield, OH
  • Developed logistic regression models and data visualizations for program evaluation

Skills

• Machine Learning
  • Physics-Informed Neural Networks (PINNs)
  • Automatic differentiation
  • Sparse regression
  • System identification
  • Bayesian hyperparameter optimization
• Data Assimilation & Modeling
  • 4D-Variational Data Assimilation (4D-Var)
  • Numerical methods
  • Dynamical systems
  • Coupled physical-biogeochemical modeling
• Programming & Scientific Computing
  • Python (PyTorch, NumPy, SciPy)
  • R, MATLAB, SQL
  • High-Performance Computing (HPC)
• Languages: Conversational Spanish

Publications

Physics-Informed Neural Networks as Differentiable Surrogates for 4D-Variational Data Assimilation

Egan, K. & Powell, B. (2026). "Physics-Informed Neural Networks as Differentiable Surrogates for 4D-Variational Data Assimilation." Journal of Advances in Modeling Earth Systems. Accepted for publication.

Automatically discovering ordinary differential equations from data with sparse regression

Egan, K., Li, W. & Carvalho, R. (2024). "Automatically discovering ordinary differential equations from data with sparse regression." Communications Physics. 7, 20.

Automatic Extraction of Ordinary Differential Equations from Data: Sparse Regression Tools for System Identification

Egan, K. (2023). "Automatic Extraction of Ordinary Differential Equations from Data: Sparse Regression Tools for System Identification." PhD thesis, Durham University.

Talks

Physics-Informed Neural Networks as Differentiable Surrogates for 4D-Variational Data Assimilation

March 25, 2026

Poster at Accelerating Research in the Age of AI (Google–UH), Honolulu, HI

Physics-Informed Neural Networks as Differentiable Surrogates for 4D-Variational Data Assimilation

February 09, 2026

Poster at Machine Learning and Dynamical Systems (MLDS 5) & Differential Equations for Data Science (DEDS 2026), Kyoto, Japan

Leveraging Physics-Informed Neural Networks for 4D-Variational Data Assimilation

June 01, 2025

Talk at CBIOMES Annual Meeting 2025, Simons Foundation, New York, NY

Building a Physics-Informed Neural Network for NPZ Modeling

March 04, 2025

Workshop at CBIOMES Workshop on Photosynthesis and Primary Production, Simons Foundation, New York, NY

Discovering Governing Equations from Data using Sparse Regression

August 01, 2021

Talk at Joint Statistical Meetings (JSM) & Symposium on Data Science & Statistics (SDSS), Virtual

Discovering Governing Equations from Data using Sparse Regression

July 01, 2021

Talk at SIAM Conference on Optimization (OP21) & Conference on Dynamical Systems (DS21), Virtual

Teaching Experience

Undergraduate Research Mentor

Workshop Instructor and Lead Organizer

Language and Culture Assistant

Math Workshop Tutor

Elementary School Teaching Assistant

Elementary Summer School Math Teacher

Software Contributions

• ARGOS (Automatic Regression for Governing Equations): Creator and maintainer of CRAN-published R package for system identification from time-series data
• PI-NPZ: Developer of PyTorch-based framework for 4D-Variational Data Assimilation using physics-informed neural networks

Grants, Workshops & Leadership Experience

• Lead Organizer & Presenter: CBIOMES Workshop on Machine Learning Methods for Time-Series Data (2026)
  Funded by a Simons Foundation Targeted Award
• Committee Member: Durham University Postgraduate Research Conference (2022)
• Chairman: Delta Tau Delta, Wittenberg University (2013–2016)